all use cases
// for business analysts

You already know the model.
Draw it, and run it.

Analysts spend their working lives modelling: entities, relationships, states, rules. Then the model becomes a document, the document becomes a backlog item, and the backlog item waits. dForge takes the diagram itself — draw your entities and relationships, import the file, and you have a working rough cut the same afternoon.

you draw this
Customer
id · name · vat
Order
id · date · total
Invoice
id · amount · status
one command
a working rough cut
entitiesviewsmenusfoldersroles

Then you and the AI finish it.

// sound familiar?

The model is right. It just never becomes software.

You can model it in an afternoon — then wait two quarters for it to exist.
The spec is precise, and what ships still isn't what you modelled.
Your ERD sits in a document nobody opens after kickoff.
Every change means another hand-off, estimate and queue.
// how dForge fits

Your diagram is the deliverable.

Draw the model, import it, finish it with AI — and hand over a running system instead of a specification.

01

Draw it the way you already do

Entities, fields and relationships — in dbdiagram.io or anything that exports DBML, or as plain SQL. This is the artefact you were producing anyway. You just stop treating it as documentation.

  • Bring DBML or SQL — the diagram you already drew.
  • Entities, fields and relationships carried over as you drew them.
  • No blank file, and no hand-off to translate your intent.
your schema → a dForge module
legacy.sql · schema.dbml · screens
model & build
dForge module — modeled
Customerid · name · vat → Orders
Orderid · date · total → Customer
Invoiceid · amount · status → Order
02

One command turns it into a module

The schema importer converts your diagram into a rough-cut dForge module: your entities and fields carried over, with views, menus, folders and roles inferred. Then you and the AI finish it — the actions, rules and reports a diagram was never able to hold.

  • Views, menus, folders and roles inferred for you.
  • You and the AI add the actions and rules a diagram can't carry.
  • It all lands as reviewable metadata, never as code you must trust.
describe it

"Track client onboarding — stages, an owner, a due date."

Client Onboardinggenerated
Client
Stage
Owner
Due
Acme Co
Kickoff
Dana
Apr 12
Northwind
Docs
Lee
Apr 18

Database, forms and roles — no developer.

03

What you hand over is a system, not a picture

A real PostgreSQL schema, generated screens, role-based access, an audit trail and reports — running, for actual users. When the business changes you change the model, not a ticket.

  • A real database, screens, permissions and an audit trail.
  • Reports people build themselves, inside their own access.
  • Change the model later — no rewrite, no queue.
security architecture
Database-per-tenant isolation
Composable row / column security
Full audit trail — every change logged
Self-host for data residency requirements
Every tool built as reviewable metadata
One access model across all modules
// your options today

Spec and wait — vs. model and run.

Spec → backlog → build
dForge
From model to something usable
A quarter, if it gets prioritised
The same afternoon
Who translates your model
A developer, reading a document
The importer, reading your file
Fidelity to what you designed
Whatever survives the hand-off
Your entities, as you drew them
Changing the model
A new ticket and an estimate
Edit the model, install again
What you hand over
A specification
A running system
Readable in a year
If someone kept the document
Always — it is the model